Discussion around managing projects and increasing the probability of success is many times focused on making things simpler. But simple only works on simple problems. Simple problems are not really that interesting if you adopt the Edwin Land view of the world.
Don't undertake a project unless it is manifestly important and nearly impossible.
But that quote has a pre-fix - Don't do anything that someone else can do. Working on easy projects means the work is easy, the problem is easy, and the solution is easy. The assumption that simple solutions are scalable to hard problems needs to be tested. Land suggests:
We can be dramatic, even theatrical; we can be persuasive; but the message we are telling must be true.
So how can we test the notion that complex and hard problems have simple solutions. Let's look at a program I am familar with and have worked a small portion of - Future Combat System (FCS).
The program is mostly canceled now - for all the right reasons. It was a bad idea, the face of combat has changed, and the program's organization and execution was wickedly complex, complicated, and hard. But that doesn't mean there are hard projects still waiting for solutions. The Affordable Care Act web site is one. It is a hard problem to integrate 50 states (some opted out), with multiple providers, on a short schedule. Having never done that before, the HHS was probably not prepared to handle the hardness of the project, let alone manage the work processes. Making changes late in the cycle was a fatal mistake from my experience in enterprise ERP. No phased delivery - you don't even need agile - just releases across the lifecycle that worked.
These are likely going to be the root cause of the problem. The project was hard but not treated as hard.
When we encounter hard problems we need to act accordingly. Suggesting there are easy solutions to hard problems is not the first response. There are no easy solutions to hard problems unless you are about to invent a new paradigm. There are very few people who can invent a new paradigm. Many making those suggestions don't work on actual hard problems, like FCS, enterprise ERP with 50 sites around the world, or integrating 50 or so ERP systems into a seamless financial reporting system for monthly status to a government agency. Problems like that. Land has advice on searching for solutions to problems like this.
We use bull's eye empiricism. We try everything, but we try the right thing first!
So what are the right things? Good question. Who gets to say what is right and what can be improved. Depends on the domain and context. If you work in a domain guided by governance, then for the most part, right is already in place. If you work for yourself or with a small group of friends right can be emergent to the needs. If the money you are spending belongs to someone else, they'll have something to say how you spend it.
In the end it comes down to this.
For every complex problem there is an answer that is clear, simple, and wrong. H. L. Mencken